Multi-epoch SMA observations of the L1448C(N) protostellar SiO jet
DDraft version November 11, 2020
Typeset using L A TEX twocolumn style in AASTeX62
Multi-epoch SMA observations of the L1448C(N) protostellar SiO jet
Tomohiro Yoshida, Tien-Hao Hsieh, Naomi Hirano, and Yusuke Aso Department of Astronomy, Kyoto University, Kitashirakawa-Oiwake-cho, Sakyo-ku, Kyoto, 606-8502, Japan Academia Sinica Institute of Astronomy and Astrophysics, 11F of ASMA Building, No.1, Sec. 4, Roosevelt Rd, Taipei 10617, Taiwan Korea Astronomy and Space Science Institute (KASI), 776 Daedeokdae-ro, Yuseong-gu, Daejeon 34055, Republic of Korea (Accepted Nov 9, 2020)
Submitted to ApJABSTRACTL1448C(N) is a young protostar in Perseus, driving an outflow and an extremely high-velocity (EHV)molecular jet. We present multi-epoch observations of SiO J = 8 −
7, CO J = 3 − ∼ . (cid:48)(cid:48)
06 yr − and ∼ . (cid:48)(cid:48)
04 yr − forthe blue- and red-shifted jet, respectively. The corresponding transverse velocities are ∼
78 km s − (blueshifted) and ∼
52 km s − (redshifted). Together with the radial velocity, we found the inclinationangle of the jets from the plane of the sky to be ∼ ◦ for the blueshifted jet and ∼ ◦ for theredfshifted jet. Given the new inclination angles, the mass-loss rate and mechanical power were refinedto be ∼ . × − M (cid:12) and ∼ . L (cid:12) , respectively. In the epoch of 2017, a new knot is detectedat the base of the redshifted jet. We found that the mass-loss rate of the new knot is three timeshigher than the averaged mass-loss rate of the redshifted jet. Besides, continuum flux has enhanced by ∼
37% between 2010 and 2017. These imply that the variation of the mass-accretion rate by a factorof ∼ ∼ −
20 yr. In addition, a knot in the downstream of theredshifted jet is found to be dimming over the three epochs.
Keywords:
Stellar jets (1607), Star formation (1569), Protostars (1302), Interstellar medium (847),Stellar winds (1636) INTRODUCTIONStudying protostellar jets/outflows helps us to under-stand the star formation process. Some protostars atan early evolutionary stage have extremely high-velocity(EHV) jets (e.g., Bachiller 1996), which may play an im-portant role for removing angular momentum and thusallowing circumstellar material to accrete onto the cen-tral stars. In addition, molecular jets are believed tobe launched from the innermost region of an accretiondisk. Although the launching mechanism is still unclear,disk accretion should provide the material to be ejectedin the form of molecular jets. Thus, studies of the jetproperties provide us a chance to explore the collapseand accretion processes in star formation (see the re-
Corresponding author: Tomohiro [email protected] view by Lee 2020). Interestingly, most molecular jetshave knotty structure (e.g., Lee 2020). The chain ofknots not only indicates the periodic ejection of the jet(e.g., Vorobyov et al. 2018) but can also be a usefultracer of the jet motion on the plane of the sky (e.g.,Girart & Acord 2001; Jhan & Lee 2015).L1448C (or L1448-mm) in the Perseus molecular com-plex ( d = 293 pc; Ortiz-Le´on et al. 2018) is a goodexample of an outflow-driving source with EHV jets.The EHV components have been observed in severalSiO transitions (e.g., Bachiller et al. 1991; Nisini et al.2007). Such SiO detections suggest molecular forma-tion via gas-phase reactions by a high mass-loss rate( ∼ − M (cid:12) yr − ) jet launched from the inside of thedust sublimation radius of the protostar (e.g., Glassgoldet al. 1991; Tabone et al. 2017; Lee 2020).A high angular resolution ( ∼ (cid:48)(cid:48) ) study in L1448C withthe Submillimeter Array (SMA; Ho et al. 2004) was a r X i v : . [ a s t r o - ph . S R ] N ov Yoshida et al. performed by Hirano et al. (2010). L1448C consistsof two sources, L1448C(N) and L1448C(S), for whichL1448C(N) is found to power the outflows with EHVjets. The two-sided jet consists of a chain of knotsthat are deflected at ∼ (cid:48)(cid:48) from the central source.The authors attributed this deflection to the preces-sion or wobbling of the disk in an unresolved binarysystem. Furthermore, the direction of the disk-jet sys-tem shows the oscillation could also occur in a singlestar system if the rotation axis is inclined to the globalmagnetic field (Machida et al. 2020). The physical pa-rameters of the outflow were derived including outflowmass, momentum, kinetic energy, mean velocity, dynam-ical timescale, mass-loss rate, momentum supply rate,and mechanical power. As a result, Hirano et al. (2010)suggested that the mass and the age of the central starare 0 . − . M (cid:12) and (4 − × yr, respectively,implying that the central star is in an extremely earlystage of evolution.However, some of the physical parameters in Hiranoet al. (2010) are inclination dependent. To derive theinclination angle, Girart & Acord (2001) observed theproper motion of the SiO clumps together with the ra-dial velocity by comparing the observations with thePlateau de Bure Interferometer (PdBI) in 1990 and theBIMA millimeter array in 1998 − ∼ ◦ . However, their analysis wasdone using observations with a low angular resolution of ∼ (cid:48)(cid:48) which is much larger than the sizes of the jet knots( ∼ (cid:48)(cid:48) ) resolved in the SMA map. Besides, the emissionfrom the jet knots is extended in the lower excitationline (SiO J = 2 − J = 8 − J = 3 − ∼ (cid:48)(cid:48) betteridentifies the proper motions of the individual knots. Inaddition, the higher excitation state of SiO J = 8 − OBSERVATIONSThe observations of 3 epochs were carried out with theSMA in 2006, 2010, and 2017. The observation dates,the array configurations, the synthetic beam sizes, andthe rms noise levels at 1 . − width are shown inTable 1. All observations contain the SiO J = 8 − J = 3 − ∼ (cid:48)(cid:48) at 345 GHz. In 2006, two pointings separated by 17 (cid:48)(cid:48) were observed, while in 2010 and 2017, only single point-ing centered at the position of L1448C(N) ( α (J2000) =3 h m s .87, δ (J2000) = 30 d m s .35) was observed.The data in 2006 have been reported in Hirano et al.(2010), and here we discuss the observations in 2010 and2017. The data in 2010 consist of extended and very ex-tended array configurations (Table 1). The data in 2010were obtained using the ASIC correlator, which dividedeach sideband of 4 GHz bandwidth into 48 “chunks”of 104 MHz width. We used the configuration that gave256 channels per chunk for the SiO line and 128 channelsper chunk for the CO line. The corresponding velocityresolution were 0.35 km s − for the SiO line and 0.7 kms − for the CO line. The data were calibrated with theMIR package (Qi 2005). The nearby quasar 3C84 wasused for amplitude and phase calibrations. The fluxcalibrators were Neptune in the very extended config-uration, and Callisto and Ganymede in the extendedconfiguration. The bandpass was calibrated by observ-ing 3C454.3. The calibrated visibility data were Fouriertansformed and CLEANed using the MIRIAD package(Sault et al. 1995). The image cubes of SiO and COwere made with a velocity interval of 1 km s − . Thesynthesized beam size of the SiO map was 0 . (cid:48)(cid:48) × . (cid:48)(cid:48) ◦ and that of the CO mapwas 0 . (cid:48)(cid:48) × . (cid:48)(cid:48)
41 with a position angle of 84 ◦ with a ro-bust weighting of 0.5.Observations of the 3rd epoch in 2017 were made inthe extended configuration using the SWARM correla-tor that covers 8 GHz bandwidth with 4 “chunks” of2 GHz width and two orthogonally polarized receivers,Rx345 and Rx400, simultaneously. Two receivers weretuned to the same frequency setting, which covered thefrequency ranges of 340.3–348.3 GHz and 356.3–364.3GHz in the lower sideband and upper sideband, respec-tively. The spectral resolution is 140 kHz across theentire band. The velocity resolution corresponds to 0.12km s − at 347 GHz. The data of each receiver werecalibrated independently using the MIR package (Qi2005). The quasars 3C84 and 3C273, and Uranus wereused as the gain calibrator, the bandpass calibrator and ulti-epoch observations of the L1448C(N) protostellar jet uv distance of > kλ for all the three epochs. The continuum mapsin 2006 and 2010 were imaged by combining the lowerand upper sideband. On the other hand, in 2017, onlythe lower sideband was used because of wider bandwidthfor the new receiver (8 GHz). Then, the images were re-grided and convolved to the same cell size and beamsize as those of the 2017 observations. The resultingbeam size of the continuum image was 0 . (cid:48)(cid:48) × . (cid:48)(cid:48)
63 witha position angle of − ◦ . After the convolution, the rmsnoise level of the 345 GHz continuum emission was 6.0mJy beam − . RESULTS3.1.
Shift of the continuum peak
In order to study the SiO proper motion, we firstlychecked the continuum source positions in the threeepochs. Figure 1 shows the continuum images and theircentral positions in the 3 epoch observations. The con-tinuum centers are obtained by fitting a 2D-Gaussian inthe image domain.The central positions and their uncertainties from thefitting are listed in Table 2. The amounts of shift of thecentral positions are much smaller than the beam size.The uncertainties of the fitting were up to ∼
11 mas inthe direction of R.A., and ∼ ∼ . (cid:48)(cid:48)
012 yr − in the direction from northwest to southeast, with a po-sition angle of ∼ ◦ .3.2. Proper Motions of the SiO knots
High-velocity SiO emission observed in 2010 is shownin Figure 2. Note that Figure 2 shows the original reso-lution image before the convolution and re-griding. TheSiO knots are found in the EHV jets ( >
40 km s − )from L1448C(N). The position angle of the EHV jet is ∼− ◦ in the blueshifted component, while it is ∼− ◦ in the redshifted component. With the high angularresolution, the SiO jet is resolved in its transverse direc-tion especially in the redshifted part. The width of thejet increases gradually toward the downstream with anopening angle of ∼ ◦ . The intrinsic width of the jet(deconvolved by the beam) is ∼ . (cid:48)(cid:48)
28 (85 au) at a pro-jected distance of 300 au from the driving source, andbecomes ∼ . (cid:48)(cid:48) ◦ and 18 ◦ for the blue- andred-shifted components, respectively so that the jets arealigned along the vertical axis. The triangles in Fig-ure 3 indicate the knot positions that were obtainedin the position-velocity diagrams (see the next para-graph). The outermost knots, BII-b and RII-b, werenot included in this analysis because those knots aremisaligned with the jet axis.To obtain the positions and radial velocities of eachknot, position velocity (PV) diagrams were used. Figure4 shows the PV diagrams of these 3 epochs along the SiOjet axis. The PV cuts are shown with the black brokenlines in Figure 2. The origins of the jets in each epochwere adopted to be the central position of the continuumemission described in section 3.1. In the high velocitypart of this diagram (faster than ±
40 km s − ), severalpeaks were found. We identified these knots in the PVdiagram by 2D-hyperboloid fitting. Table 3 shows thecentral position and the radial velocity of each knot.As a result, in Figure 3, the slopes of the lines that con-nect the knots in three epochs reflect the proper motionof the knots. Assuming that each knot moves at a con-stant velocity, we performed a linear fitting in the spacedomain. Figure 5 shows a plot of the derived proper mo-tions of the knots versus the distances of the knots fromthe central source. The uncertainty comes from the lin-ear fitting given the errors of the positions derived fromthe 2D-hyperboloid fitting. The averages of the propermotions are ∼ . (cid:48)(cid:48)
06 yr − for the blueshifted side and ∼ . (cid:48)(cid:48)
04 yr − for the redshifted side. The correspondingtransverse velocities are ∼
78 km s − and ∼
52 km s − for the blue- and red-shifted jet, respectively.The transverse versus radial velocities of the knots areplotted in Figure 6. The dashed lines show the results oflinear fitting to the knots on each side; this fitting givesthe inclination angles measured with respect to the plane Yoshida et al.
Table 1.
Properties of each observationDate Array configuration Beam size Weighting Noise level (K) a . (cid:48)(cid:48) × . (cid:48)(cid:48)
84 uniform 1.912/25/2006 Extended9/17/2010 Extended 0 . (cid:48)(cid:48) × . (cid:48)(cid:48)
41 robust=0.5 1.57/19/2010 Very Extended11/10/2017 Extended 0 . (cid:48)(cid:48) × . (cid:48)(cid:48)
69 robust=0.5 2.3 a The rms noise levels of the SiO ( J = 8 −
7) maps at the channel width of 1 km s − measured after being concolved to the same beam. D e c ( J ) Figure 1.
Continuum maps made using the visibility data with the uv distance >
70 k λ in three epochs. The first contour is5 σ and the contour interval is 5 σ where 1 σ = 6.0 mJy beam − . The red stars indicate the central positions obtained by fittinga Gaussian. The green grids are the same in all panels. Table 2.
Central positions of the continuum emissionEpoch Central position (J2000)2006 3 h m . ± . d m . ± . h m . ± . d m . ± . h m . ± . d m . ± . of the sky to be (34 ± ◦ and (46 ± ◦ for the blue-and red-shifted jet, respectively. The derived inclinationangles are consistent with those estimated through themodel of infrared scattered light from the outflow cavity(Tobin et al. 2007). They are also consistent with theinclination angle at the base of the redshifted jet derivedfrom the proper motions of the H O maser spots ( ∼ ◦ at a distance of 293 pc, Hirota et al. 2011). Besides,the 3D velocity of each jet knot was measured by usingits transverse and radial velocities. The representative3D velocities of the jets ( ∼ ± − and ∼ ± − for the blue- and red-shifted side, respectively)were derived by averaging that of every knot on eachside. Figure 7 shows the 3D velocity versus the 3D distanceto the central source in the first epoch; the 3D distancehere is measured by using the inclination angles. The 3Dvelocity of the knot tends to increase as the distance in-creases. This apparent acceleration seen in Figure 7 canbe caused by decreasing density along the jet. Indeed,the opening angle of ∼ ◦ (Figure 2) of the jet suggeststhat the jet decreases its density as it travels. Anotherpossibility is the variation in the ejection velocity of thejet itself. If the ejection velocity gradually decreasesas a function of time, the measured velocity becomeslarger as the distance from the central source increases.In fact, it is clear that the newly ejected BI-a knot hasa higher velocity than that of BI-b, suggesting that theejection velocity can vary. The decrease/increase of ejec-tion velocity can be interpreted as a consequence of ex-pansion/contraction of the jet launching radius possiblydue to the variation in mass-accretion rate.In addition, the BII-a knot is ∼
50% faster than theRII-a knot. Such asymmetry in jet velocities are oftenseen in the optical/infrared jets from Class II sources(e.g., Hirth et al. 1994). Theoretically, it is proposed ulti-epoch observations of the L1448C(N) protostellar jet RA offset (") D e c o ff s e t ( " ) Figure 2.
High-velocity SiO emission observed in 2010with the original angular resolution. The velocity ranges are∆ V = ± −
70 km s − with respect to the systemic velocity( v LSR ∼ − ). The contours start at 3 σ and increase in5 σ intervals with 1 σ = 0 .
23 Jy beam − . The black brokenlines indicate the PV cuts with position angle of ∼ ◦ forthe blueshifted part and ∼ ◦ for the redshifted part. Theblack and green stars indicate the positions of the protostars,L1448C(N) and L1448C(S), respectively. that the asymmetry occurs in the MHD disk winds ifeach side of the disk has different magnetic lever armand/or launch radius (Ferreira et al. 2006). Dyda et al.(2015) argued that the magnetic field lines in the densedisk oscillate periodically, producing the field line in oneside to be more perpendicular to the disk (reducing out-flow) and another side more inclined (favor to produceoutflow). Another possibility is that the asymmetry isa consequence of different mass loading on the oppositesides of the disk in the framework of X-wind model (Liu& Shang 2012). Such a velocity asymmetry has alsobeen observed in the jet of Class 0 source L1157-mm o ff s e t ( " ) SouthNorth
BI-aBI-bBI-cBII-aBII-bRI-aRI-bRI-cRII-aRII-b
Figure 3.
Comparison of the SiO moment 0 maps (∆ V = ± −
70 km s − ) in 3 epochs. The images of the blue-and redshifted jets were rotated counterclockwise by 25 ◦ and18 ◦ , respectively, so that the jet axes align with the verticalline. The contours start at 3 σ and increase in 3 σ intervalswith 1 σ = 0 .
73 Jy beam − . The ellipses on the right topand bottom corner are the synthesized beams for the blue-and red-shifted side, respectively. The triangles are the peakpositions of each knot in the PV diagrams (Figure 4). BI-a/RI-a and BI-b/RI-b are not separated in the moment 0map, while they are different knots in Figure 4. (Podio et al. 2016), which suggests that the mechanismsimilar to the asymmetric jets in Class II sources alsoworks in the Class 0 cases.3.3. Physical parameters of the EHV jets
Because of the unknown SiO abundance, the phys-ical parameters of the EHV jet were estimated usingthe CO flux measured in the velocity ranges of ∆ V = ± − from the systemic velocity ( v LSR ∼ − ). In the velocity ranges of ∆ V = ± −
50 km s − , the CO emission from the jets is contami-nated by the outflow shells associated with L1448C(N)(see Section 3.4). Thus, in order to exclude the con- Yoshida et al.
RI-a
RI-b
RI-c
RII-a
BI-a
BI-b
BI-c
BII-a
Figure 4.
PV diagrams of the SiO jets in the 3 epochs. The contours start at 2 σ and increase in 3 σ intervals where1 σ = 0 .
13 Jy beam − . The positions of each knot are marked with black crosses. The systemic velocity of this source, v sys , is ∼ − . Table 3.
Properties of knotsBI-a BI-b BI-c BII-aOffset | v − v sys | Offset | v − v sys | Offset | v − v sys | Offset | v − v sys | (arcsec) (km s − ) (arcsec) (km s − ) (arcsec) (km s − ) (arcsec) (km s − )2006 0.62 66.0 1.87 44.7 5.33 59.0 10.9 59.82010 0.94 58.2 1.93 42.3 5.36 59.5 11.0 59.72017 1.45 55.9 2.18 43.4 5.86 61.2 11.7 58.3RI-a RI-b RI-c RII-aOffset | v − v sys | Offset | v − v sys | Offset | v − v sys | Offset | v − v sys | (arcsec) (km s − ) (arcsec) (km s − ) (arcsec) (km s − ) (arcsec) (km s − )2006 0.93 49.5 2.63 53.6 4.89 59.2 9.83 62.62010 1.02 50.0 2.71 54.6 5.03 60.3 10.0 60.82017 1.12 48.7 2.96 53.7 5.48 59.5 10.4 62.6 ulti-epoch observations of the L1448C(N) protostellar jet p r o p e r m o t i o n ( " y e a r − ) Figure 5.
Proper motions of the knots and the distancefrom the central source. The broken lines show the averagedvalue representing the red- and blue-shifted sides of the jet. sys | (km s −1 )020406080100120 t r a n s v e r s e v e l o c i t y ( k m s − ) BI-aBI-b BI-cBII-aRI-a RI-bRI-c RII-a
Figure 6.
Radial versus transverse velocities of each knot.The uncertainty of the transverse velocity originates from thelinear fitting given the position errors of the 2D-hyperboloidfitting, and that of the radial velocity comes from the errorof the 2D-hyperboloid fitting itself. The dashed lines meanthe fitted line for the knots on each side. The angles fromthe the plain of the sky (y-axis) are 34 ◦ for the blue line, and46 ◦ for the red line. tamination of the shell components, we narrowed theregion in the ranges of ∆ V = ± −
50 km s − . Weused the image observed in 2006 with the compact andextended configurations, because it recovered 80–100%of the CO flux at ∆ V > ±
20 km s − . On the otherhand, the CO images observed in 2010 and 2017 with-out compact configuration suffer from the missing fluxof up to 90% as compared to previous single-dish obser- D v e l o c i t y ( k m / s ) Figure 7.
3D velocities of each knot versus its 3D distancefrom the central star in the first epoch. The dashed linesindicate the linear fits to the data points except for the BI-aknot.
Table 4.
Dynamical parameters of the EHV jet fromL1448C(N)Parameters Blue RedMass a ( M (cid:12) ) 4.3 × − × − Momentum a ( M (cid:12) km s − ) 0.043 0.031Kinetic energy b (erg) 4.4 × × Mean velocity (km s − ) 98 ± ± a ( M (cid:12) yr − ) 1.1 × − × − Mechanical power b ( L (cid:12) ) 0.93 0.34 a The typical uncertainty is estimated to be ∼
10 %. b The typical uncertainty is estimated to be ∼
20 %. vations (Nisini et al. 2000) because the CO emission isspatially extended even in the EHV velocity ranges. Weassumed that the CO emission from the jet is opticallythin, and that the excitation condition of the CO followsthe local thermal equilibrium (LTE). The mean atomicweight was adopted to be 1.41. The CO abundance of4 × − is assumed. This high value is proposed by thechemical model of protostellar winds in Glassgold et al.(1991). We assume an excitation temperature of 100 Kas used in Hirano et al. (2010) for comparison.The derived physical parameters of the jets are givenin Table 4. The mass of the EHV jet is ∼ × − M (cid:12) .The uncertainty of the mass estimation mainly arisesfrom that of the absolute flux calibration. Since the SiOflux values at the middle knots of the jet measured in the Yoshida et al.
RA offset (") D e c o ff s e t ( " ) Figure 8. CO J = 3 − ± −
40 km s − with respect to the systemic velocity.The contour interval is ∼ − km s − (3 σ ) withthe lowest contour level of ∼ − km s − (3 σ ).The black and green stars indicate the central sources of theL1448C(N) and L1448C(S), respectively. The orange crossmarks the position of RII-b knot. three epochs show good agreement with each other (seeSection 3.6), the measured flux is accurate to ∼ l/v , where l is the 3D length of thejet and v is the mean 3D velocity of the jet. Here, theprojected length of the blueshifted CO jet was measuredto be ∼ (cid:48)(cid:48) and, that of the redshifted CO jet was ∼ (cid:48)(cid:48) .(see Figure 5 in Hirano et al. 2010). The mean velocitiesof the jets were given in Section 3.2. Thus, the dynami-cal timescale of the blueshifted jet and the redshifted jetwere estimated to be ∼
390 yr and ∼
600 yr, respectively.As a result, we obtained the total mass-loss rate of thejet to be ∼ × − M (cid:12) yr − . The total mechanicalpower of the jet was found to be ∼ L (cid:12) .3.4. Physical parameters of the outflow shells
The CO outflow shells associated with L1448C(N) arealso detected (Figure 8). The extended shell-like struc-tures are only seen in the velocity ranges of ∆ V ∼± −
50 km s − . The physical parameters of the outflowshells were calculated using CO data in 2006. Even with the compact configuration, the missing flux in this veloc-ity range is significant. By comparing to the single-dishCO J = 3 − ∼ is 10 − , and the meanatomic weight is 1.41. Note that we adapt the standardabundance because we consider that most of the gas inthe lower velocity outflow shells is the swept-up ambientmaterial. We also assumed the excitation temperatureof ∼
40 K following Hirano et al. (2010). It was estimatedfrom the peak brightness temperature of CO around thesystemic velocity, assuming that the CO emission at thisvelocity is optically thick. To calculate the dynamicalparameters, the inclination angles of the outflows areassumed to be the same as those of the jets.Table 5 shows the calculated physical parameters ofthe outflow shells with and without missing flux correc-tion. After the correction, the total mass of the outflowshells was estimated to be ∼ × − M (cid:12) . The dy-namical timescale of the outflow is defined as l/v , where l and v are the 3D length and the 3D velocity of theoutflow. The projected length of the CO outflows onboth sides was estimated to be ∼ (cid:48)(cid:48) . Thus, the dy-namical timescale of the blueshifted shell is calculatedto be ∼ ∼ ∼ × − M (cid:12) km s − yr − and ∼ . L (cid:12) with miss-ing flux correction. Since the mechanical power of thetwo-sided jet is ∼ . L (cid:12) , the jet has sufficient power todrive the outflow shell. In addition, the momentum sup-ply rate of the jet is ∼ . × − M (cid:12) km s − yr − forboth sides of the jet, which is comparable to the valueof the outflow shell. Therefore, the shells can be drivenby the jet, although we cannot exclude the possibilitythat the outflow and the jet are launched independently(e.g. Machida et al. 2008; Machida 2014).3.5. Appearance of a new knot
Figure 9 shows the SiO moment 0 maps in the RIregion overlaid on those of CO in three epochs. The in-tegrated ranges of these maps are ∆ V = 61 −
70 km s − for both SiO and CO lines. A newly appeared jet knotis seen in the 2017 image at a position near to RI-a;this component is also seen in the PV diagram of 2017with a velocity much higher than that of RI-a (Figure4). The new knot is located at a distance of ∼ . (cid:48)(cid:48)
77 fromthe central star. Although the new knot is clearly seen ulti-epoch observations of the L1448C(N) protostellar jet Table 5.
Dynamical parameters of the L1448C(N) outflow shellBlue RedParameters uncorrected a corrected b uncorrected a corrected b Mass d ( M (cid:12) ) 2.6 × − × − × − × − Momentum d ( M (cid:12) km s − ) 0.094 0.16 0.12 0.23Kinetic energy e (erg) 4.9 × × × × Momentum supply rate c,d ( M (cid:12) km s − yr − ) 3.1 × − × − × − × − Mechanical power c,e ( L (cid:12) ) 0.13 0.17 0.14 0.21 a The effect of missing flux is not corrected. b The effect of missing flux is corrected channel by channel in comparison with the single-dish observations. c The dynamical timescale is assumed to be ∼ ∼ d The typical uncertainty is estimated to be ∼
10 %. e The typical uncertainty is estimated to be ∼
20 %. in the high-velocity moment 0 map, there is no clearboundary between it and RI-a in the PV diagram (Fig-ure 4). However, toward the position of the new knot,the CO/SiO intensity increases significantly in the widevelocity range of ∆ V ∼ −
70 km s − .We define the mass of the new knot as the mass in-creased between 2006 and 2017. Since this new knotis also seen in the CO map in the same velocity range,the mass of the new knot was derived using the COmap. In order to estimate the increased mass, we firstlyderived the total mass of the RI-a and RI-b knots con-sidering the proper motions. The velocity range wastaken to be ∆ V = 51 −
70 km s − . The same physicalconditions used for the EHV jet (see Section 3.3) wereassumed in the estimation. The total mass of the RI-a + RI-b knots is estimated to be ∼ . × − M (cid:12) in2006, ∼ . × − M (cid:12) in 2010, and ∼ . × − M (cid:12) in 2017. Then, the mass of the new knot is estimatedto be ∼ × − M (cid:12) by subtracting the average of the2006 and 2010 values from the one of 2017. If we assumethat the 3D velocity of this knot is same as the averagevelocity of the redshifted jet, ∼
78 km s − , this knothas a momentum of ∼ × − M (cid:12) km s − and a me-chanical power of ∼ . L (cid:12) . If the new knot is ejectedalong the axis of the redshifted jet, the distance betweenthis knot and the central source is ∼
330 au. With the3D velocity of ∼
78 km s − , the dynamical timescale ofthe knot is estimated to be ∼
20 yr. Thus, the aver-aged mass-loss rate of the new knot is estimated to be ∼ . × − M (cid:12) yr − . However, our assumption ofthe excitation temperature might be underestimated insuch a newly shocked region. In addition, the dynamical Table 6.
Dynamical parameters of the newknot Parameters the new knotMass a ( M (cid:12) ) 3 . × − Momentum b ( M (cid:12) km s − ) 2 . × − Mean velocity (km s − ) 78 ± b ( M (cid:12) yr − ) 1 . × − Mechanical power b ( L (cid:12) ) 0.85 a The typical uncertainty is estimated to be ∼ b The typical uncertainty is estimated to be ∼ timescale is the upper limit for duration time of the jeteruption. Therefore, the mass-loss rate derived here isconsidered as the lower limit.3.6. Brightness variability of the jet knots
The line profiles at the emission peaks of the SiO knotsare shown in Figure 10. The line profiles in the threeepochs are consistent toward the BI-a, RI-b, RI-c, andRII-a knots. In the BI-b and BI-c knots, the peak in-tensity remains similar, although the intensity of thelower velocity tail measured in 2017 is lower than thoseof 2006 and 2010. The peak intensity of the RI-a knotincreased significantly, from ∼
15 K in 2006 to ∼
25 K in2017. This likely originates from the appearance of thenew knot. On the other hand, the knots downstream, —0
Yoshida et al.
RI-b
New knot
RI-c
Figure 9.
SiO map (contours) and CO map (color scale) in three epochs. Both the SiO and CO emission were integratedover the velocity range of ∆ V ∼ −
70 km s − . The contours are drawn every ∼ − km s − (3 σ ) with the lowestcontour at ∼ − km s − (3 σ ). The noise level of the CO map is ∼ − km s − . White star marks thecentral source. i.e. BII-b and RII-b — have dimmed, especially RII-b;in 2006, the peak brightness temperature of the RII-bknot was ∼
15 K, while, it became ∼ J = 3 − ∼ (cid:48)(cid:48) .Since the observations of the three epochs have beendone in different array configurations, the effect of spa-tial filltering to the observed intensity might be differentin each epoch. To study the effect of the spatial filtering,we have simulated the observations of the redshifted jetusing the same uv sampling as that of the observationsof each epoch with the MIRIAD package. The inputmodel was composed of four deconvolved 2D-Gaussiansobtained by fitting to the SiO moment 0 map in 2006(Figure 3). The modeled images were convolved andre-grided to the image of the first epoch. Simulation re-sults (Figure 12) show that the brightness of the knotssuch as RI-c and RII-a is not always constant, implyingthat different uv sampling could introduce an intensityvariation of ∼ uv sampling (Figure 12). In addition, the in-tensity dimming observed in the RII-b knot from ∼
15K to ∼ ∼ uv sampling. Thus, we consider Table 7.
Gaussian fit results for the compact componentEpoch Peak amplitude (Jy) FWHM ( (cid:48)(cid:48) )2006 0.31 0.372010 0.32 0.392017 0.41 0.35 that the dimming of the RII-b knot in SiO should bereal rather than an effect of the spatial filtering.3.7.
Variability of the continuum emission
Figure 13 shows the visibility amplitudes of the con-tinuum emission in the three epochs as a function of uv distance. This plot was made using the sideband withthe common frequency. Each data point was obtainedby azimuthally averaging the data in a 5 k λ bin. Thecontinuum emission can be decomposed into the com-pact and extended components, using the boundary at ∼
70 k λ . The visibility amplitude profile of each epochwas fit by two Gaussian components. Due to the lackof the data points in the shortest baseline range, theGaussian fitting is not well constrained in the extendedcomponent. On the other hand, the missing flux doesnot affect the results of the compact component. TheGaussian fit results of the compact component in thethree epochs are listed in Table 7. The flux values ofthe compact component in 2006 and 2010 are 0.30 Jyand 0.32 Jy, respectively, and are consistent with eachother. However, the flux in 2017 is 0.41 Jy and is ∼ ulti-epoch observations of the L1448C(N) protostellar jet -20-50-800102030 B r i g h t n e ss T e m p e r a t u r e ( K ) BI-a -20-50-80
BI-b -20-50-80
BI-c -20-50-80
BII-a -20-50-80
BII-b
RI-a
RI-b v − v sys (km s −1 )RI-c RII-a
RII-b
Figure 10.
Line profiles of SiO J = 8 − − )relative to the systemic velocity, and the y-axis is the brightness temperature (K). The black solid line, the blue dash dottedline, and the red dashed line shows the line profile in 2006, 2010 and 2017, respectively. -20-50-800102030 B r i g h t n e ss T e m p e r a t u r e ( K ) BI-a -20-50-80
BI-b -20-50-80
BI-c -20-50-80
BII-a -20-50-80
BII-b
RI-a
RI-b v − v sys (km s −1 )RI-c RII-a
RII-b
Figure 11.
Line profiles of CO J = 3 − − ) relative to the systemic velocity, and the y-axis is the brightness temperature (K). The black solid line, theblue dash dotted line, and the red dashed line shows the line profile in 2006, 2010 and 2017, respectively. Yoshida et al. model ←RII-b
RA offset (") D e c o ff s e t ( " ) Figure 12.
Results of the simulated observations using the same uv coverage of the observations at each epoch. The leftmostpanel shows the input model. The intensity is normalized by the maximum value in the input model. The contours start from0.1 and drawn every 0.2. ulti-epoch observations of the L1448C(N) protostellar jet Figure 13.
Visibility amplitude of the continuum emissionas a function of the uv distance. The black, blue, and reddots are the data points of the 2006, 2010, and 2017 obser-vations, respectively. The profile of each epoch was fittedwith two Gaussian components. The dotted lines indicatethe Gaussian fit results to the compact component, and thesolid curves are the total amplitude of the two components. seen in the literature (e.g., Liu et al. 2018; Yoo et al.2017), implying a change of the mass-accretion rate. DISCUSSION4.1.
Mass-loss rate and mass-accretion rate
The mass-loss rate by the jet is believed to be relatedto the mass accretion rate to the central star. The mass-accretion rate ˙ M acc is expressed by˙ M acc ∼ L bol R (cid:63) GM (cid:63) , (1)where L bol is the bolometric luminosity of the centralsource, R (cid:63) is the stellar radius, G is the gravitationalconstant, and M (cid:63) is the stellar mass. The bolomet-ric luminosity of the central source is measured to be L bol ∼ L (cid:12) (at the updated distance of 293 pc) byTobin et al. (2016). The stellar mass derived from theanalysis of the Keplerian motion of the circumstellardisk was estimated to be ∼ . − . M (cid:12) , assumingthe averaged inclination angle of ∼ ◦ (Maret et al.2020). If we assume a stellar radius of ∼ R (cid:12) from the-oretical studies (e.g., Stahler 1988), the mass-accretionrate is estimated to be ∼ (2 − × − M (cid:12) yr − . Ifthis is the case, the ratio of the mass-loss rate to themass-accretion rate becomes ˙ M loss / ˙ M acc ∼ . − .
9. Ifwe adopt a smaller stellar mass of ∼ . M (cid:12) , the ratiois ∼ .
3, which is consistent with the value predictedby the X-wind model (e.g., Shu et al. 2000). Alterna-tively, this high ˙ M loss / ˙ M acc implies that the averaged mass-accretion rate in the star-forming process is higherthan the current value derived from the observations. Inaddition, assuming a constant mass-accretion rate andstellar mass of ∼ . M (cid:12) , the age of the central star isestimated to be ∼ . × yr. This value is longer thanthe dynamical timescale of the larger scale ( ∼ (cid:48) ) out-flow (Bachiller et al. 1990; Gomez-Ruiz et al. 2013) bya factor of ∼
10. As a result, both the high ˙ M loss / ˙ M acc and age of the star suggest that the mass-accretion rateand mass-ejection rate are variable. This result is com-monly found in protostellar outflows driven by low-massYSOs. (Dunham et al. 2010; Lee et al. 2013; Takahashiet al. 2013; Hsieh et al. 2016).4.2. The new knot
The mass of the new knot, ∼ × − M (cid:12) , is ∼ ∼ × − M (cid:12) . In addition, the angular sepa-ration of < (cid:48)(cid:48) between the new knot and the next one,RI-a, is much smaller than the typical separation of theadjacent knots. These imply that the new knot is a “sub-knot (smaller knotty structure)” of RI-a. Indeed, simi-lar hierarchical knotty structure is also observed in theHH211 protostellar jet in the Perseus; the sub-knots inthe HH211 jet have an angular separation of ∼ . (cid:48)(cid:48)
75 (seeFigure 5 in Lee et al. 2009). The new knot in L1448C(N)could be a similar structure as the sub-knots in HH211.The new knot has a higher velocity ( v ∼
80 km s − )than the precedent knots. This feature is similar to thatof the BI-a knot (Figure 4), which suggests that the newknot could be the counterpart of the BI-a knot. This isconsistent with the fact that each knot likely has itscounterpart (Figures 3, 4 and 7).The appearance of the new knot suggests that themass-accretion rate varies in a short time scale, andincreased ∼
20 yr ago (the dynamical timescale of theknot). The averaged mass-loss rate of the new knot of ∼ . × − M (cid:12) yr − is ∼ ∼ ∼
3. Such a luminosity change of the centralsource warms the circumstellar disk by T ∝ L . for ablackbody (see Yoo et al. 2017). If this is the case, theflux density at 350 GHz is expected to be enhanced by ∼
30% after the bolometric luminosity increases by afactor of ∼
3. Indeed, as mentioned in Section 3.7, thecontinuum flux has increased by ∼
37% between 20104
Yoshida et al. and 2017. Taking into account the dynamical timescaleof the new knot of ∼
20 yr, this continuum enhancementis unlikely to be the origin of the new knot. However,the enhancement in both mass-loss rate and continuumflux implies that the mass accretion in L1448C(N) variesin a short timescale of ∼ −
20 yr.4.3.
Dimming of the knots in the downstream
In order to explain the dimming of the RII-b knot inthe downstream of the redshifted jet, we examine thepossibility of changes in density and temperature due tothe knot expansion (Figure 10). We adopt the initial H density of ∼ . × cm − and kinetic temperature of ∼
400 K, which are obtained from the multi-transitionSiO (from J = 2 − J = 11 −
10) single-dish observa-tions of Nisini et al. (2007). Using the non-LTE radia-tive transfer code RADEX (Van der Tak et al. 2007), wefind that an SiO column density of ∼ . × cm − isrequired to reproduce the peak brightness temperatureof ∼
15 K of RII-b (Figure 10) given the line width of∆ V ∼
10 km s − . We then assume an isotropical adia-batic expansion of the knot that increases the volume bya factor of ∼
2. After the expansion, the gas tempera-ture, the H density, and the SiO column density become ∼
270 K, ∼ . × cm − , and ∼ . × cm − , re-spectively. This results in a brightness temperature ofSiO down to ∼ . ∼
30 km s − , assuming theknot was initially concentrated within ∼ (cid:48)(cid:48) (Figure 3).This velocity is broadly consistent with the line widthof the knot in SiO/CO within a factor of ∼ − ∼ × cm − at300 K (Yang et al. 2010).However, it is unclear why only the RII-b knot is dim-ming, but the other knots are constant. A possible hy-pothesis is that the RII-b knot is affected by the outflowfrom another protostar, L1448C(S) (Figure 2). Indeed,Hirano et al. (2010) found that the outflows driven byL1448C(N) and L1448C(S) may interact with each other(Figure 8) in the redshifted side. This implies that theredshifted jet from the L1448C(N) also can be affectedby the outflow from L1448C(S). Since the RII-b knotis located near the outflow cavity of L1448C(S) on theplane of the sky (Figure 8), the expansion of the RII-bknot can be caused by the interaction with the outflowfrom L1448C(S).Another possibility is the change of ejection direction.As seen in Figures 2 and 3, the jet is deflected at theposition in between RI and RII, and BI and BII. This de-flection suggests a change of ejection direction caused by the precession or wobbling of the circumstellar disk. Insuch a case, the RII-b knot is traveling in a low-densityregion dissipated by the former jet. On the contrary, theRII-a knot may be plowing into the fresh denser mate-rial. If this is the case, the RII-b knot can expand easierthan the knots in the current jet axis.In addition, the SiO abundance in the knot can bevariable. Assuming an optically thin SiO line and fixingthe physical conditions in the LTE, the dimming of theline suggests that the abundance of the SiO moleculewas reduced by a factor of two. The abundance ofthe SiO molecule is supposed to drop via reaction ofSiO + OH → SiO + H, since OH is abundant in theshocked gas. The rate coefficient of this reaction is pro-posed to be k ∼ (2 − × − cm s − (Le Teuff et al.2000; Gusdorf et al. 2008; Langer & Glassgold 1990).Thus, the corresponding reaction timescale is estimatedto be t = 1 /k ∼ − ∼
10 yr,which is much shorter than the reaction timescale. CONCLUSIONSWe compared 3 epochs—2006, 2010 and 2017—ofSMA observations to study L1448C(N) protostellar jetsin SiO J = 8 − J = 3 − ∼ . (cid:48)(cid:48)
06 yr − and ∼ . (cid:48)(cid:48)
04 yr − for the blue-and red-shifted side, respectively. The correspond-ing transverse velocities are ∼
78 km s − (blue)and ∼
52 km s − (red). The inclination angleof the EHV jet was estimated by comparing thetransverse and radial velocities of the jet knots.The inclination angle (from the plane of the sky)of the blueshifted jet is ∼ (34 ± ◦ and that ofthe redshifted jet is ∼ (46 ± ◦ . The jet axis is ∼ ◦ more inclined than the previous estimation.The mean velocity of the EHV jet is estimated tobe ∼ (98 ±
4) km s − for the blueshifted jet and ∼ (78 ±
1) km s − for the redshifted jet.2. The mass-loss rate of L1448C(N) is estimated tobe ∼ . × − M (cid:12) yr − . The ratio of the mass-loss rate to the mass-accretion rate is found tobe ∼ . − .
9, depending on the current stellarmass. The lower-end value is broadly consistentwith the theoretical predictions. The higher-end iscommonly found in protostellar jets/outflows andis interpreted by a variation in mass-accretion rate.The mechanical power of L1448C(N) is refined tobe ∼ . L (cid:12) . ulti-epoch observations of the L1448C(N) protostellar jet ∼ . × − M (cid:12) . The lower limit of the mass-loss rate ofthe new knot is estimated to be ∼ . × − M (cid:12) yr − . If this is the case, the mass-accretion ratebecomes ∼ . × − M (cid:12) , which is higher than theaveraged mass-accretion rate by a factor of ∼ ∼
37% higher than that in 2006and 2010. The enhancement in both mass-loss rateand continuum flux implies that the mass accre-tion in L1448C(N) varies in a short timescale of ∼ −
20 yr.5. The SiO flux of the RII-b knot in the downstreamof the redshifted jet decreased by ∼
50% in the last ∼
10 yr. This dimming of SiO can be explainedby the change of physical condition caused by theexpansion of the knot.We would like to thank the SMA staff for theirhelp during these observations. The SMA is a joint project between the Smithsonian Astrophysical Obser-vatory and the Academia Sinica Institute of Astronomyand Astrophysics and is funded by the Smithsonian In-stitution and the Academia Sinica. We are grateful toour referee, Dr. Rafael Bachiller, for the helpful com-ments. We also thank Dr. Anthony Moraghan for care-fully proofreading the manuscript. N.H. acknowledgesa grant from the Ministry of Science and Technology(MoST) of Taiwan (MoST 108-2112-M-001-017, MoST109-2112-M-001-023).
Facility:
SMA (Ho et al. 2004)
Software:
MIR (Qi 2005), CASA (McMullin et al.2007), MIRIAD (Sault et al. 1995), , RADEX (Van der Taket al. 2007), Astropy (Astropy Collaboration et al. 2013;Price-Whelan et al. 2018), APLpy (Robitaille & Bressert2012)REFERENCES